COMMENTARY a Chromomeric Model for Nuclear and Chromosome Structure

Journal of Cell Science 108, 2927-2935 (1995) 2927 Printed in Great Britain © The Company of Biologists Limited 1995

COMMENTARY A chromomeric model for nuclear and chromosome structure

Peter R. Cook CRC Nuclear Structure and Function Research Group, Sir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford OX1 3RE, UK e-mail: [email protected]

SUMMARY

The basic structural elements of chromatin and chromo- into mitosis, increased adhesiveness between nucleosomes somes are reviewed. Then a model involving only three and between factories drives a ‘sticky-end’ aggregation to architectural motifs, nucleosomes, chromatin loops and the most compact and stable structure, a cylinder of nucle- transcription factories/chromomeres, is presented. Loops osomes around an axial chromomeric core. are tied through transcription factors and RNA polymerases to factories during interphase and to the remnants of those factories, chromomeres, during mitosis. On entry Key words: chromomere, nucleoskeleton, RNA polymerase, factory

INTRODUCTION meric model will be presented; it helps explain how the interphase fibre is converted into a mitotic chromosome and why a ‘We are thus brought, finally, to one of the most fundamental chromosome has its characteristic shape. conceptions of cytology and genetics, namely, that the A number of inter-related factors complicate analysis of spireme-threads are linear aggregates of much smaller self-per- chromosome structure. First, their tight packing and high petuating bodies, aligned in single series, and in definite order.’ density makes it difficult to visualize their interior and limits As chromosomes condense, these bodies, or chromomeres, access of probes like antibodies. Second, chromatin is poised may ‘aggregate or fuse to form larger ones’ and ‘be compound in a metastable state so that even small changes in tonicity bodies having perhaps a definite internal architecture.’ cause it to aggregate into an unworkable mess. Therefore These quotations are taken from the concluding section on unphysiological conditions are used in almost all isolation pro- chromosome structure in E. B. Wilson’s classic review of cell cedures; for example, conventional cytological preparations biology published in 1928 (Wilson, 1928); they summarize a are made by hypotonically swelling cells, followed by half-century of careful microscopic analyses of chromosomes fixation/extraction in methanol-acetic acid and an explosive of many different organisms. Strings of chromomeres were spreading at an air-liquid interface. Third, chromosomal sub- traced from interphase as they condensed, split and segregated structures have sizes below the resolution of the light micro- during mitosis, to decondense in the daughter cells. Whilst the scope, necessitating electron microscopy which brings compound term spireme-thread (spirema: a thing wound or problems associated with preserving structure in vacuo. coiled, a skein) expresses an ambivalence as to whether the Therefore, studies on isolates made using more physiological string condensed spirally, the reader is left in no doubt that conditions or on rapidly-frozen specimens will be stressed. chromosomes are built around chromomeres. Despite the wealth of evidence available to Wilson, chromomeric models fell out of favour, largely because they did BASIC FEATURES OF CHROMOSOME not lead to a plausible explanation of why chromosomes have ORGANIZATION the shape that they have. Furthermore, Wilson presumed that some vestige of the chromomere persisted during interphase, Condensation is the hallmark of mitosis. Why, then, does the but no suitable structure could be found. As a result, most chromatin fibre not condense into the most compact form, a current models involve some form of helical coiling or ill- sphere? Why are chromatids cylindrical, and not spherical? defined interactions between DNA and/or proteins to account for chromosome structure (Fig. 1). However, a good candidate Helical hierarchies for the interphase counterpart of the mitotic chromomere has It is a truism that extended biological structures are helical at now been uncovered. Here, the basic structural elements of the molecular level, but not at the macro-molecular level; actin chromatin and chromosomes that any model must take into filaments and DNA are helical, but cells and trees, which are account will first be reviewed and then an up-dated chromo- nevertheless made of those helical molecular assemblies, are 2928 P. R. Cook not. This truism arises because assembly of molecular sub- has, in 1 mM monovalent ion, the canonical ‘beads-on-a- units into a non-helical structure requires that each sub-unit is string’ structure visible by electron microscopy (Thoma et al., positioned next to its neighbours with no axial rotation; even 1979). As the ionic strength is raised progressively, the string the slightest adds up over many sub-units to give a helix. folds into a series of helices with a fairly constant pitch but Therefore we might expect a hierarchy of helices in a chro- increasing numbers of nucleosomes per turn; at ~60 mM, the mosome: the double helix first being coiled into a nucleosome resulting ‘solenoid’ is ~25 nm wide, with 6-8 nucleosomes/turn (~11 nm diam.), nucleosomes into solenoids (~30 nm diam.), and an 11 nm pitch. But both right- and left-handed solenoids solenoids into chromatid fibres (225-250 nm diam.), and so on are seen in the original micrographs, implying some disorder. to give a cylinder (Fig. 1E). At each level in the hierarchy, the Moreover, others interpret rather similar structures in terms of sense of coiling and pitch between monomers would be strings of ‘superbeads’ (Hozier et al., 1977; Strätling et al., precisely defined. According to this view, one might solve (in 1978). Increasing the salt concentration further has little effect a crystallographic sense) the structure of a mitotic chromosome until chromatin precipitates at ~100 mM, so these experiments and expect some remnants of the hierarchy to be retained into provide no direct evidence that solenoids exist at a physiolog- interphase. However, when biological structures become as ically-relevant salt concentration (i.e. ~150 mM monovalent long as cells, they also become sufficiently flexible that other ion). forces, like those involved in growth, can overcome the weaker There is also no evidence for solenoids in vitrified sections forces maintaining helicity. A long disordered structure may of metaphase CHO or HeLa cells that have been neither fixed well relax into a helix when the other forces are removed in nor stained (McDowall et al., 1986). Chromosomes have a vitro, so the crucial question is whether any coiling then seen homogeneous ‘grainy’ texture and optical diffractograms are actually pre-existed in vivo. best explained by a compact association of 11 nm particles There is no reason to believe that the DNA coils seen in interacting in a manner akin to molecules in a liquid. It is nucleosomal crystals do not exist in vivo, but the evidence for unlikely that solenoids did exist but were missed, because coiling at the next level of organization is controversial. microtubules with the same mass/unit-length can be seen in the Chromatin released from nuclei by digestion with a nuclease cytoplasm. Electron tomography of interphase nuclei embedded at low temperatures also reveals ribbons of zig- zagging nucleosomes and not solenoids; variable lengths of linker DNA seem to enter and exit individual nucleosomes at various angles in the disordered structure, instead of the more precise arrangement expected of a solenoid (Horowitz et al., 1994). There is also controversy over the coiling of the chromone- mal fibre (Manton, 1950): gyres can vary with growth tem- perature in Tradescantia (Swanson, 1942), they may appear constant in number but variable in sense in hypotonically- swollen human chromosomes (Ohnuki, 1968), of opposite sense in a minority of the sister chromatids of isolated HeLa ‘scaffolds’ (Boy de la Tour and Laemmli, 1988), or right- handed in polytene Drosophila chromosomes (Hochstrasser and Sedat, 1987). Coupled with the gap in the hierarchy at the level of the solenoid, such variability smacks more of variability in growth rate or in the conditions at the time of analysis than of some general structural principle. Moreover, an attractive feature of coiled-coil models is that progressive rotation can progressively condense the coil, but chromatids condense without rotation; human vimentin genes tend to retain the same external position on sister chromatids, irrespective of chromatid length (Baumgartner et al., 1991). 30 nm fibres Fig. 1. Some models for chromosome structure during G1 phase and mitosis. Possible G1 structures include (A) a randomly-folded fibre, A ~30 nm fibre is widely believed to be another architectural loops attached (B) to the lamina (e.g. Gerace and Burke, 1988), motif but, again, there is little agreement on its precise dimen- (C) to chromomeres (e.g. Zatsepina et al., 1989) or scaffolds (e.g. sions, the amount of DNA/unit-length and its existence in vivo Paulson and Laemmli, 1977), (D) to proteins bound at specific sites (e.g. van Holde, 1988). Recent careful work suggests that a true on the fibre, or (E) a coiled-coil (e.g. Rattner, 1992). After biological variation underlies this controversy (Woodcock, duplicating DNA and dissolving skeletons, mitotic structure might be 1994). It seems that 30 nm fibres are found in transcription- dictated by interactions (F) within randomly-folded chromatids (e.g. ally-inactive cells like chicken erythrocytes with condensed DuPraw 1966), (G) between chromomeres or scaffolding elements chromatin and ‘long’ (i.e. >210 bp) nucleosomal repeats, but (e.g. Manuelidis, 1990), or (H) between gyres that tighten a coiled- coil (e.g. Bak et al., 1977). Hybrid models (not shown) might involve not in active cells with ‘typical’ repeats of 160-200 bp. For coiled-coils and loops attached to a skeleton or combinations of example, 30 nm fibres are clearly seen in frozen-hydrated radial loops and helical folding (e.g. Rattner and Lin, 1985). sections of nuclei of starfish sperm that were swimming in sea Chromatids could be arranged symmetrically, as in G and H. water up to the moment of cryo-immobilization (Woodcock, Nuclear and chromosome structure 2929

1994). In contrast, most active nuclei contain no such fibres their molecular ties are defined and the topoisomerase is strate- visible in vitrified sections (Dubochet et al., 1988). gically placed to decatenate and/or condense loops by altering Most chromatin is transcriptionally inert, even in ‘active’ supercoiling (Saitoh and Laemmli, 1994). But scaffolds retain nuclei, so we might expect most to form 30 nm fibres, but it their morphology when topoisomerase II is removed (Hirano does not. Why? Presumably, transcription of a few sequences and Mitchison, 1993; Swedlow et al., 1993) and five out of six somehow unfolds fibres throughout the nucleus. of their loops are created during isolation (Jackson et al., 1990). Looping has also been inferred from the rate at which Chromatin loops and underlying sub-structures nucleases solubilize the chromatin of isolated nuclei (Igó- Another recurrent structural motif involves attaching the Kemenes and Zachau, 1977). Cutting an unlooped fibre should chromatin fibre in loops to a sub-structure. Apparently decisive first release two long fragments that are then shortened, but the evidence is provided by phase-contrast microscopy of unfixed expected long fragments are not seen; rather, the kinetics are ‘lampbrush’ chromosomes of amphibians (Callan, 1977). Such consistent with two cuts releasing one short fragment from a structures are prepared by puncturing an oocyte, extruding the 75 kbp loop. But again, long unlooped fibres do aggregate as nucleus and then removing its envelope in 0.1 M KCl/NaCl; nuclei are isolated (Jackson et al., 1990). In principle, tightly- the jelly-like clump of chromatin slowly disperses to reveal attached sequences can be identified by detaching most DNA loops attached to a chromomeric axis (Macgregor and Varley, with a nuclease and pelleting the sub-structure, but pellets of 1988). But there is no hint of individual chromomeres or loops cages, matrices and scaffolds contain different sequences and in electron micrographs of intact nuclei; instead the chromatin it is not clear which reflects a structure found in vivo (Cook, appears as one granular aggregate. Therefore, it is quite 1988). possible that transcription units/loops are stripped off the If any consensus can be drawn from these conflicting results granules during dispersal. Indeed, possible intermediates in obtained using hypo- and hyper-tonic conditions, it is probably such a process, small granules, are often seen scattered around that the chromatin fibre is looped, but how, and to what, is not loops (e.g. Mott and Callan, 1975). clear. Indeed, one model even involves attaching loops to the Supercoiling provides additional evidence for looping. one skeleton that we are confident does exist, the peripheral Supercoils are lost spontaneously from linear DNA as the lamina (Fig. 1B; Gerace and Burke, 1988); chromatin’s high molecule can spin about its axis; therefore, the existence of concentration close to lamin proteins and its affinity for them supercoiling implies that DNA is tied down (i.e. looped) to are consistent with this (Traub and Shoemann, 1994). prevent rotation. ‘Nucleoids’ isolated by lysing interphase cells in >1 M NaCl sediment in gradients containing ethidium like Chromosome bands superhelical DNA (Cook and Brazell, 1975); electron After various treatments, certain reagents stain some chromoso- microscopy also reveals naked superhelical loops attached to mal regions more intensely to give chromosome-specific bands: a residual ‘cage’ (Jackson et al., 1984). The contour length of G/Q bands are AT-rich, contain facultative heterochromatin and such loops remains unchanged during mitosis, so their basic are late-replicating; R bands are GC-rich, contain ~80% known structure must persist (Jackson et al., 1984). Significantly, genes plus many alu sequences, and are early-replicating; C supercoiling and the cage are both lost as transcriptionally- bands generally contain tandem repeats of centromeric DNA active chick erythroblasts mature into inert erythrocytes, (Craig and Bickmore, 1993). The total number of bands depends pointing to a connection between looping and transcription on the degree of resolution but, to a broad approximation, up to (Cook and Brazell, 1976). Whilst such loops could be created 1,250 can be seen in prematurely-condensed chromosomes, by artifactually on isolation, the total number actually falls ‘replication’ banding and in prophase (Hameister and Sperling, slightly, and the remaining ones are attached through the same 1984; Drouin et al., 1990). As cells progress into mitosis, indi- points as those seen in isolates made using isotonic buffers vidual R bands fuse more quickly and in greater numbers than (Jackson et al., 1990). (Note that unrestrained supercoils are G bands to reduce the total (Drouin et al., 1991). Trypsin found only locally in eukaryotic chromatin; Jupe et al., 1993.) treatment is usually used to generate G bands, but a similar Nuclear ‘matrices’ are superficially similar to nucleoids but pattern of slightly thicker regions can be seen in untreated human are isolated using both hypo- and hyper-tonic steps (reviewed metaphase spreads by atomic force microscopy (Musio et al., by Getzenberg et al., 1991). It is often assumed that they are 1994), in prematurely-condensed chromosomes (Gollin et al., associated with superhelical loops, but the DNase treatment 1984) and in nucleoids (Mullinger and Johnson, 1980). generally used during isolation ensures that they are not. (‘In Although homologous chromosomes are similarly shaped, they situ matrices’ with superhelical DNA are prepared more like rarely have exactly the same banding patterns or lengths, and nucleoids; Vogelstein et al., 1981.) Moreover, during the initial this is why progress on the automation of karyotyping is so slow. stages of isolation, the number of loops (and so attachments) Even so, any model for chromosome structure should be able to first increases and then decreases (Jackson et al., 1990), account for these bands. making it difficult to be certain which, if any, of those seen finally actually existed in vivo. Chromosomal proteins An axial ‘scaffold’ associated with (unsupercoiled) loops Despite considerable effort, the list of proteins implicated in was originally observed by electron microscopy of histone- chromosome structure is limited. It includes histone H1 (which depleted chromosomes (Fig. 1G; Paulson and Laemmli, 1977). is hyper-phosphorylated during mitosis; Bradbury et al., 1974), Specific scaffold-associated regions (SARs) are attached to the RNA polymerases I and II (Matsui et al., 1979), topoisomerase major protein in the isolate, topoisomerase II (Mirkovitch et II (see above), and XCAP-C and E (Hirano and Mitchison, al., 1984; Earnshaw and Heck, 1985). Scaffolding models are 1994). The latter were identified after mixing sperm chromatin attractive because isolated scaffolds look like chromosomes, in a mitotic extract from Xenopus eggs and then spinning the 2930 P. R. Cook assembled chromosomes through sucrose; the two proteins, encapsulating cells in agarose microbeads (50-150 µm diam.) which are homologous to a scaffold protein (Saitoh et al., before permeabilizing them in a ‘physiological’ buffer (Fig. 2; 1994), remained bound. They are recruited to the chromomeric Jackson et al., 1988). The chromatin, now protected by core as chromosomes assemble, whilst antibodies against agarose, does not aggregate and can be pipetted freely; it is XCAP-C inhibit this process. accessible to probes like antibodies and enzymes, its DNA is It is often assumed that the residual proteins (e.g. topoiso- intact and polymerases retain their activity. The basic struc- merases, XCAPs) bound to an extracted chromosome (e.g. a tural features of HeLa nuclei, including a strong candidate for scaffold) necessarily determine its shape, but this is not so: the chromomere of interphase, have been analyzed using shape survives extraction with a strong detergent like dodecyl ‘physiological’ conditions and the approaches illustrated in sulphate which removes all the proteins discussed above Fig. 2. (Cook, 1984). Partial deproteinization of DNA packed as The results can be summarized as follows. The lamins, tightly as a mitotic chromosome will inevitably generate a which are members of the intermediate-filament family of tangle that roughly reflects the original shape, and then the proteins and which form an exo-skeleton that underpins the mere association of any residual proteins with the tangle tells nuclear membrane (Gerace and Burke, 1988), are also part of us little about their structural role. Moreover, disrupting the an internal nucleoskeleton (Jackson and Cook, 1988; Hozák et function of any proteins involved in condensation, whether al., 1995). Chromatin loops with an average contour length of they be topoisomerases, XCAPs, histones or kinases, using 86 kbp are attached to this skeleton; this length does not change antibodies (as in the case of XCAP-C) or inhibitors would be during mitosis (Jackson et al., 1990). Surprisingly, loops, expected to affect the condensed shape. Therefore there whether part of the natural chromosome or a transfected remains little good evidence for the (currently-fashionable) ‘minichromosome’, are attached through promoters/enhancers role invoked for topoisomerase II and the XCAPs in main- and transcribed sequences to polymerizing complexes on the taining shape, although both are clearly required to generate skeleton (Jackson and Cook, 1985, 1993). The active poly- shape. merases that mediate attachment are concentrated into discrete ‘foci’ (Jackson et al., 1993; Wansink et al., 1993). (The term ‘polymerase’ is used to describe the large cluster of polypep- STUDIES USING ‘PHYSIOLOGICAL’ CONDITIONS tides in the active complex.) The foci have variable sizes and shapes, with the largest apparently formed by fusion; this Arguments whether isolated structures are generated artifactu- makes counting difficult, but in a HeLa cell there are 300-500 ally by the hyper- or hypo-tonic conditions used are best bright ones and several thousand in all, or roughly the number countered by the use of more physiological conditions. Fortu- of mitotic bands (F. Iborra, A. Pombo, D. A. Jackson and P. nately, the practical problems caused by chromatin aggrega- R. Cook, unpublished results). The bright ones contain >50 tion at an isotonic salt concentration can be overcome by active polymerases, plus associated templates and components

Fig. 2. A procedure for analyzing chromatin structure using ‘physiological’ conditions. (A) HeLa cells are (B) encapsulated in an agarose bead (dotted surroundings). (C) After permeabilization, the cytoskeleton, lamina, internal nucleoskeleton (all in brown), associated transcription factory (red oval) and DNA loop (blue line) covered with nucleosomes (green circles) all become accessible to molecular probes. (D) Added endonucleases can now diffuse through the agarose and cut chromatin loops (arrows) so that (E) most chromatin can be removed electrophoretically. (F) Skeletons, whether in the nucleus or cytoplasm, are best visualized by electron microscopy of thick sections. (Redrawn from Jackson and Cook, 1985.) This procedure has been used to characterize: (i) An internal lamin-containing nucleoskeleton, once obscuring chromatin is removed (Jackson and Cook, 1988; Hozák et al., 1995). (ii) The contour length of loops, from the average length and percentage of remaining DNA fragments (if fragment length is 8.6 kbp and 10% remains, contour length is 8.6×1/(10/100) = 86 kbp). It does not change during mitosis, so the molecular ties holding loops persist (Jackson et al., 1990). (iii) Sequences remaining after elution; they are mainly promoters, enhancers and transcribed sequences, implying that engaged polymerases, which can still ‘run-on’ along residual fragments, mediate attachment to the skeleton (Jackson and Cook, 1985, 1993). (iv) Sites of transcription. Permeabilized cells (either before or after cutting and elution) are allowed to make RNA in the presence of Br-UTP, and then sites (i.e. factories) containing the incorporated analogue are immunolabelled using antibodies against Br- RNA (Jackson et al., 1993; Wansink et al., 1993). Nuclear and chromosome structure 2931 of the splicing apparatus, so they are called transcription surface area and so the number of polymerases accessible to ‘factories’ by analogy with the replication factories that are promoters. When nucleoli disassemble during mitosis, most also fixed to the skeleton and contain all the machinery nucleolar components disperse (Hernandez-Verdun and necessary to duplicate >20 replicons simultaneously Gautier, 1994) but, remarkably, all polymerase I and most of (Nakamura et al., 1986; Hozák et al., 1993, 1994b). the transcription factor, UBF, remain bound as the remnants of the centres aggregate into the nucleolar organizing regions on the chromosomes (Scheer and Rose, 1984; Roussel et al., NUCLEOLAR TRANSCRIPTION FACTORIES 1993). Therefore, the polymerase I factories of interphase are directly related to mitotic chromomeres. Extra-nucleolar transcription factories (containing RNA poly- It is then attractive to suppose that factories containing polymerases II and III) have not yet been well-characterized by merases II and III are built similarly. Increased transcription electron microscopy, but their nucleolar counterparts contain- would disaggregate cores, increasing surface area and the ing RNA polymerase I have (Spector, 1993). Nucleoli contain number of accessible polymerases. When transcription ceases several ‘fibrillar centres’ surrounded by a ‘dense fibrillar during mitosis, most core material would disperse and the rest component’; in turn, these are embedded in a ‘granular would aggregate with other remnants to give the chromomeres component’. Fibrillar centres, which are located at nodes on the skeleton, store the required enzymes and transcription seems to occur as rDNA slides through polymerases on their surface. Nascent rRNA is then extruded into the dense fibrillar component and, after termination, it moves to the granular component to complete its maturation. Therefore these factories, which are several hundred nanometers in diameter, contain the active machinery on the surface of storage cores, the fibrillar centres (Hozák et al., 1994a). There is detailed information on the development of these factories. Each fibrillar centre is often associated with one, or a few, active ribosomal cistrons and the total number of centres is directly related to the rate of rRNA synthesis. For example, the ~234 in a fibroblast fall to ~156 on serum-starvation (Jordan and McGovern, 1981) and the ~9 in a peripheral blood lymphocyte rise to ~80 as it is stimulated to divide (Haaf et al., 1991). In other words, increasing transcription increases

Fig. 3. (A-D) Cell sorting through differential adhesiveness and (E- G) an analogous ‘sticky-end’ aggregation and sorting during prophase. (A-D) When two cell types (green and red) are (A) mixed, (B) they aggregate into the most compact form, a sphere. (C) As contacts created by random movement between the more-adhesive (red) cells persist for longer, they clump internally. (D) Clumps of more-adhesive cells continue to aggregate until the less-adhesive cells surround a few large clumps, although sorting is rarely perfect. (Redrawn from Steinberg, 1964.) (E-G): (E) A string of less- adhesive nucleosomes (green) is attached to two more-adhesive factories (red). (F) As factories touch (arrow) during condensation, they stick together; large factories are probably the first to fuse. Displaced loops create a higher density in the plane of contact; for geometrical reasons this density is ~5× and ~10× higher if 86-kbp loops stretch 350 nm from factories of 50 and 25 nm diameter, respectively. (G) As the touching halves of the two factories fuse into a cylinder (arrowhead), this central density increases further by one-third. The nucleosomal concentration is now much higher around the middle of the cylinder than at an end; this enhances nucleosomal aggregation and ensures that the next factory to fuse will do so at a more-accessible end. Moreover, the next factory is generally close by and tethered through a short inter-factory loop, as it split from its neighbour earlier during interphase when a larger loop attached to generate two smaller loops, one of which became the inter-factory loop. Additional factories now bind at the ends, elongating the cylinder. These geometrical considerations, and the principles illustrated in A-D, ensure that chromatids are cylindrical (and not spherical) and that loops initially incorporated into the ‘wrong’ cylinder (usually a sister chromatid) will sort into the ‘right’ one. 2932 P. R. Cook of mitosis. Groups of small factories rich in genes would creates contacts and those between the more-adhesive cells condense into Geimsa-light bands. And as the RNA poly- tend to persist; eventually a sphere of less-cohesive cells forms merases and transcription factors that tie the chromatin into around a core of the more-adhesive ones (Steinberg, 1964; loops remain bound, those loops would retain their contour Armstrong, 1989). This model has been tested directly length. (Steinberg and Takeichi, 1994). L cells are not normally very adherent, but become more so if they express (after transfec- tion of the appropriate cDNA) the adhesion molecule, A DIGRESSION ON CELLULAR ADHESION cadherin, on their surface; mixtures expressing different amounts of cadherin sort into low-expressers surrounding a How randomly-arranged chromatin condenses into a mitotic core of high-expressers. cylinder is one of the central issues that must be addressed by We can model interphase chromosomes, which are any model. The cylinder even survives translocation into a confined to discrete nuclear regions (Engh et al., 1992), as foreign environment, for example when several Mbp of S. strings of less-adhesive nucleosomes running between, and pombe DNA are integrated into a mouse cell (McManus et al., looping from, more-adhesive factories (Fig. 3E-G). In 1994). Clearly, species-specific interactions cannot be mitosis, increased adhesiveness drives ‘sticky-end’ aggrega- involved. If such cylinders are not built using helices or highly tion to the most compact and stable structure, a cylinder of specific interactions, we must look to other means exploited by nature to generate simple higher-order structures. Fig. 4. A model for the structure of a human chromosome. Upper: When disaggregated neural cells from frog embryos are structure during G1 phase. Transcription factories are located at randomly intermixed, they aggregate into the most compact nodes on a lamin-containing endo-skeleton and their number and size form, a sphere, and then sort into a semblance of the original depends upon transcriptional activity. If there are 1,250, each might tissue, with ectoderm surrounding the endoderm (Fig. 3; be ~50 nm diameter, contain ~25 active RNA polymerases and be Townes and Holtfreter, 1955). Many different cells will sort in associated with ~56 chromatin loops (usually, but not invariably, this way and the process can be modelled if cells differ only derived from one chromatid) with a range of contour lengths (i.e. 5 in their adhesiveness for one another. Then random movement sets of 11 loops centred around ~7.5, 50, 75, 100 and 175 kbp, respectively; average 86 kbp). ~10% chromatin then lies within ~8.6 kbp of a factory, so genes within it are close enough to polymerases to attach and be transcribed; the remainder is too remote and condenses on to the lamina or nucleolus as heterochromatin. Increased transcription generates more, smaller, factories each associated with fewer, shorter, loops. The enlargement (right) shows a loop attached through a transcription unit and a promoter/enhancer to a polymerizing complex and transcription factor (red ovals). (In transcriptionally-inactive regions, transcription factors are the sole molecular ties.) A transcript (wavy red line) is extruded as the template slides through the left-hand complex; this template movement ‘opens’ adjacent chromatin. The DNA duplex winds around nucleosomes (green circles) in the loop; variations in linker length (±2 bp) and entry-exit angle (±15¡) generate an irregular zig- zagging fibre (Horowitz et al., 1994) that extends, on average, ~350 nm from a factory. Lower: mitotic structure. Human chromosome 16, which contains ~100 Mbp DNA (3% of genome), is modelled, assuming each chromatid is 3,400×800 nm. (The dimensions of ‘native’ chromosomes vary significantly, depending on condensation.) On entry into mitosis, skeletons depolymerize, transcription ceases and nucleosomes plus ~38 factory remnants aggregate and then sort into a cylindrical nucleosomal ‘cloud’ surrounding a chromomeric axis, as described in Fig. 3E-G. Despite these rearrangements, loops retain their attachments and contour lengths. A typical factory/chromomere, plus an average loop extended on each side, fit within the width; longer loops fold back on themselves and are the major determinants of width (enlargement, right). Factories associated with proportionally more shorter loops (and active genes) give slightly narrower cylinders; examples include (i) Geimsa-light bands (5% thinner), (ii) several Mbp of yeast DNA translocated into a mouse genome (the closely-spaced yeast genes are transcribed during interphase and so will be in short loops; McManus et al., 1994) and (iii) active rDNA loops which give the ‘secondary constriction’ of the nucleolar organizing region (Robert- Fortel et al., 1993). The design principles are, almost certainly, further modified in other specialized chromosomal regions; for example, the several Mbp of non-transcribed α1-satellite probably condense into a spherical centromeric ‘chromomere’ through interactions involving CENP-B, rather than polymerases or transcription factors (Yoda et al., 1992). Nuclear and chromosome structure 2933 nucleosomes around a chromomeric core in which inter- CONCLUSIONS nucleosomal and inter-chromomere interactions are maximized. Then, many different loop-chromomere, nucleo- This model lies centrally within the historical tradition sum- some-nucleosome and chromomere-chromomere interactions marized by Wilson’s statement that heads this piece. It is also maintain chromosomal shape, rather than the few typical of a minimalist one, requiring only three basic structural motifs scaffolding models. in the organizational hierarchy, the nucleosome, chromatin loop and factory/chromomere, to explain why nuclei and chromosomes have the shapes they have. It involves one major A MODEL assumption: interphase transcription foci/factories are the pre- cursors of mitotic chromomeres, which is supported by the These considerations lead to a model involving three funda- established relationship of nucleolar foci/factories with the mental levels of organization: nucleosomes, loops and tran- mitotic nucleolar organizing regions. The interactions involved scription factories/chromomeres. Loops are tied through tran- are either numerous and non-specific (i.e. between nucleo- scription factors or RNA polymerases to factories during somes and between chromomeres) or fewer but more specific interphase and to chromomeres, which are the remnants of (i.e. between polymerases/transcription factors and transcrip- those factories, during mitosis. Each level is characterized by tion units). They will inevitably generate both helical and dis- flexibility without having a fixed structure. ordered structures depending on how smoothly condensation During interphase, a zig-zagging nucleosomal ribbon is occurs. The structural role suggested for RNA polymerase as looped to factories strung along an internal lamin-containing one loop tie is perhaps surprising, but the enzyme plays just nucleoskeleton (Fig. 4, top); the ribbon slides past attached such a role in looping bacterial DNA; indeed, the cluster of polymerases, changing its contour length from one moment to polymerases at the core of the bacterial nucleoid provides the the next so it cannot revert to the heterochromatic 30 nm fibre prototype for a chromomere (Krawiec and Riley, 1990). The found in inert cells. Increased transcription counteracts the persistence of all ties through mitosis inevitably means that tendency of factories to fuse and generates more smaller transcriptional patterns will be inherited by daughter cells. factories to which more genes are attached in shorter loops. Importantly, the best evidence for the model is derived from During mitosis, the skeleton depolymerizes, transcription studies on intact or cryo-fixed cells, or those permeabilized in ceases (Shermoen and O’Farrell, 1991) and proteins (including ‘physiological’ buffers. histone H1) are phosphorylated, increasing adhesiveness Wilson’s ‘self-perpetuating bodies’ are then the organizers between factories and between nucleosomes. Chromosome of nuclear structure. But they also organize function; their condensation would begin in heterochromatin and proceed RNA polymerases ensure that they are transcription centres, bidirectionally (Hiraoka et al., 1989) to give cylindrical strings and, by nucleating the formation of processing and replication of chromomeres that first fuse into one large cylinder per sites (Spector, 1993; Hassan et al., 1994), they are also inti- (entangled) chromatid pair, which then splits into two, from an mately involved in other, vital, nuclear functions. end toward a centromere (Sumner, 1991), as individual chromatids sort out. Longer strings would give longer cylinders of I thank my many colleagues (especially Dean Jackson and Guy similar width. Each cylinder would contain nucleosomes Houlsby) for helpful discussions, and The Cancer Research Campaign arranged around a disordered chromomeric axis (Fig. 4B) and The Wellcome Trust for support. which might be helical locally where a run of uniformly-sized factories/chromomeres condensed at a constant rate, or where large ones were so positioned that they forced the appropriate REFERENCES bending (Sorsa, 1986). The hypotonic treatment used by Armstrong, P. B. (1989). Cell sorting out; the self-assembly of tissues in vitro. Ohnuki (1968) could relax the structure into a more complete Crit. Rev. Biochem. Mol. Biol. 24, 119-149. helix. In different metaphases, regions rich in transcribed genes Bak, A. L., Zeuthen, J. and Crick, F. H. C. (1977). Higher-order structure of (in short loops) associated with (small) factories will condense the mitotic chromosomes. Proc. Nat. Acad. Sci. USA 74, 1595-1599. to form Geimsa-light bands in roughly the same places, but Baumgartner, M., Dutrillaux, B., Lemieux, N., Lilienbaum, A., Paulin, D. and Viegas-Pequignot, E. (1991). Genes occupy a fixed and symmetrical individual genes do not occupy fixed positions; rather, certain position on sister chromatids. Cell 64, 761-766. factories (and associated genes) might tend to end up on the Boy de la Tour, E. and Laemmli, U. K. (1988). The metaphase scaffold is outsides of a pair of chromatids, perhaps due to their size helically folded: sister chromatids have predominantly opposite helical relative to their neighbours, whilst others end up more inter- handedness. Cell 55, 937-944. nally, as is the case (Baumgartner et al., 1991). Whilst cells in Bradbury, E. M., Inglis, R. J. and Matthews, H. R. (1974). Control of cell division by very lysine rich histone (F1) phosphorylation. Nature 247, 257- different tissues have transcription patterns which are generally 261. similar, differences in detail will probably be reflected by Callan, H. G. (1977). Lampbrush chromosomes. Proc. R. Soc. London, Ser. B. small, but nevertheless detectable, differences in the fine 214, 417-448. site? J. Cell Sci. 90, 1-6. Cook, P. R. and Brazell, I. A. (1975). Supercoils in human DNA. J. Cell Sci. structure of the banding pattern in mitosis. On entry into G1, 19, 261-279. the decondensing chromatids would be confined to discrete Cook, P. R. and Brazell, I. A. (1976). Conformational constraints in nuclear nuclear domains by their neighbours. Simultaneously, the DNA. J. Cell Sci. 22, 287-302. appropriate intermediate filaments would repolymerize Cook, P. R. (1984). A general method for isolating intact nuclear DNA. EMBO between decondensing chromomeres, around the nuclear J. 3, 1837-1842. periphery and throughout the cytoplasm to form the internal Cook, P. R. (1988). The nucleoskeleton: artefact, passive framework or active Craig, J. M. and Bickmore, W. A. (1993). Chromosome bands - flavours to nucleoskeleton, lamina and cytoskeleton that together integrate savour. BioEssays 15, 349-354. cellular space. Drouin, R., Lemieux, N. and Richler, C.-L. (1990). Analysis of DNA 2934 P. R. Cook

replication during S-phase by means of dynamic chromosome banding at preparing cyto- and nucleo-skeletons and associated chromatin. J. Cell Sci. high resolution. Chromosoma 99, 273-280. 90, 365-378. Drouin, R., Lemieux, N. and Richler, C.-L. (1991). Chromosome Jackson, D. A., Dickinson, P. and Cook, P. R. (1990). The size of chromatin condensation from prophase to late metaphase: relationship to chromosome loops in HeLa cells. EMBO J. 9, 567-571. bands and their replication time. Cytogenet. Cell Genet. 57, 91-99. Jackson, D. A. and Cook, P. R. (1993). Transcriptionally-active Dubochet, J., Adrian, M., Chang, J. J., Homo, J. C., Lepault, J., McDowall, minichromosomes are attached transiently in nuclei through transcription A. W. and Schultz, P. (1988). Cryo-electron microscopy of vitrified units. J. Cell Sci. 105, 1143-1150. specimens. Q. Rev. Biophys. 21, 129-228. Jackson, D. A., Hassan, A. B., Errington, R. J. and Cook, P. R. (1993). DuPraw, E. J. (1966). Evidence for a ‘folded-fibre’ organization in human Visualization of focal sites of transcription within human nuclei. EMBO J. chromosomes. Nature 209, 577-581. 12, 1059-1065. Earnshaw, W. C. and Heck, M. M. S. (1985). Localization of topoisomerase Jordan, E. G. and McGovern, J. H. (1981). The quantitative relationship of II in mitotic chromosomes. J. Cell Biol. 100, 1716-1725. the fibrillar centres and other nucleolar components to changes in growth Engh, G. van den, Sachs, R. and Trask, B. J. (1992). Estimating genomic conditions, serum deprivation and low doses of actinomycin D in cultured distance from DNA sequence location in cell nuclei by a random walk model. diploid human fibroblasts (strain MRC-5). J. Cell Sci. 52, 373-389. Science 257, 1410-1412. Jupe, E. R., Sinden, R. R. and Cartwright, I. L. (1993). Stably maintained Gerace, L. and Burke, B. (1988). Functional organization of the nuclear microdomain of localized unrestrained supercoiling at a Drosophila heat envelope. Annu. Rev. Cell Biol. 4, 335-374. shock gene locus. EMBO J. 12, 1067-1075. Getzenberg, R. H., Pienta, K. J., Ward, W. S. and Coffey, D. S. (1991). Krawiec, S. and Riley, M. (1990). Organization of the bacterial chromosome. Nuclear structure and the three-dimensional organization of DNA. J. Cell. Microbiol. Rev. 54, 502-539. Biochem. 47, 289-299. Macgregor, H. C. and Varley, J. B. (1988). Working With Animal Gollin, S. M., Wray, W., Hanks, S. K., Hittelman, W. N. and Rao, P. N. Chromosomes. 2nd edition. John Wiley, Chichester, UK. (1984). The ultrastructural organization of prematurely condensed Manton, I. (1950). The spiral structure of chromosomes. Biol. Rev. 25, 486- chromosomes. J. Cell Sci. Suppl. 1, 203-221. 508. Haaf, T., Hayman, D. L. and Schmid, M. (1991). Quantitative determination McDowall, A. W., Smith, J. M. and Dubochet, J. (1986). Cryo-electron of rDNA transcription units in vertebrate cells. Exp. Cell Res. 193, 78-86. microscopy of vitrified chromosomes in situ. EMBO J. 5, 1395-1986. Hameister, H. and Sperling, K. (1984). Description of a chromosome Manuelidis, L. (1990). A view of interphase chromosomes. Science 250, 1533- replication unit in individual prematurely condensed human S-phase 1540. chromosomes. Chromosoma 90, 389-393. Matsui, S., Weinfeld, H. and Sandberg, A. (1979). Quantitative conservation Hassan, A. B., Errington, R. J., White, N. S., Jackson, D. A. and Cook. P. R. of chromatin-bound RNA polymerases I and II in mitosis: implications for (1994). Replication and transcription sites are colocalized in human cells. J. chromosome structure. J. Cell Biol. 80, 451-464. Cell Sci. 107, 425-434. McManus, J., Perry, P., Sumner, A. T., Wright, D. M., Thomson, E. J., Hernandez-Verdun, D. and Gautier, T. (1994). The chromosome periphery Allshire, R. C., Hastie, N. D. and Bickmore, W. A. (1994). Unusual chromosome structure of fission yeast DNA in mouse cells. J. Cell Sci. 107, during mitosis. BioEssays 16, 179-185. 469-486. Hirano, T. and Mitchison, T. J. (1993). Topoisomerase II does not play a Mirkovitch, J., Mirault, M.-E. and Laemmli, U. K. (1984). Organisation of scaffolding role in the organization of mitotic chromosomes assembled in the higher-order chromatin loop: specific DNA attachment sites on nuclear Xenopus extracts. J. Cell Biol. 120, 601-612. scaffold. Cell 39, 223-232. Hirano, T. and Mitchison, T. J. (1994). A heterodimeric coiled-coil protein Mott, M. R. and Callan, H. G. (1975). An electron microscope study of the required for mitotic chromosome condensation in vitro. Cell 79, 449-458. lampbrush chromosomes of the newt Triturus cristatus. J. Cell Sci. 17, 241- Hiraoka, Y., Minden, J. S., Swedlow, J. R., Sedat, J. W. and Agard, D. A. 261. (1989). Focal points for chromosome condensation and decondensation Mullinger, A. M. and Johnson, R. T. (1980). Packing DNA into revealed by three-dimensional in vivo time-lapse microscopy. Nature 342, chromosomes. J. Cell Sci. 46, 61-86. 293-296. Musio, A., Mariani, T., Frediani, C., Sbrana, I. and Ascoli, C. (1994). Hochstrasser, M. and Sedat, J. W. (1987). Three-dimensional organization of Longitudinal patterns similar to G-banding in untreated human Drosophila melanogaster interphase nuclei. I. Tissue-specific aspects of chromosomes: evidence from atomic force microscopy. Chromosoma 103, polytene nuclear architecture. J. Cell Biol. 104, 1455-1470. 225-229. Horowitz, R. A., Agard, D. A., Sedat, J. W. and Woodcock, C. L. (1994). Nakamura, H., Morita, T. and Sato, C. (1986). Structural organisation of The three-dimensional architecture of chromatin in situ: electron replicon domains during DNA synthetic phase in the mammalian nucleus. tomography reveals fibres composed of a continuously variable zig-zag Exp. Cell Res. 165, 291-297. nucleosomal ribbon. J. Cell Biol. 125, 1-10. Ohnuki, Y. (1968). Structure of chromosomes. I. Morphological studies of the Hozák, P., Hassan, A. B., Jackson, D. A. and Cook, P. R. (1993). spiral structure of human somatic chromosomes. Chromosoma 25, 402-428. Visualization of replication factories attached to a nucleoskeleton. Cell 73, Paulson, J. R. and Laemmli, U. K. (1977). The structure of histone-depleted 361-373. metaphase chromosomes. Cell 12, 817-828. Hozák, P., Cook, P. R., Schöfer, C., Mosgöller, W. and Wachtler, F. Rattner, J. B. and Lin, C. C. (1985). Radial loops and helical coils coexist in (1994a). Site of transcription of ribosomal RNA and intra-nucleolar structure metaphase chromosomes. Cell 42, 291-296. in HeLa cells. J. Cell Sci. 107, 639-648. Rattner, J. B. (1992). Integrating chromosome structure with function. Hozák, P., Jackson, D. A. and Cook, P. R. (1994b). Replication factories and Chromosoma 101, 259-264. nuclear bodies: the ultrastructural characterization of replication sites during Robert-Fortel, I., Junéra, H. R., Géraud, G. and Hernandez-Verdun, D. the cell cycle. J. Cell Sci. 107, 2191-2202. (1993). Three-dimensional organization of the ribosomal genes and Ag- Hozák, P., Sasseville, A. M.-J., Raymond, R. and Cook. P. R. (1995). Lamin NOR proteins during interphase and mitosis in PtK1 cells studied by confocal proteins form an internal nucleoskeleton as well as a peripheral lamina in microscopy. Chromosoma 102, 146-157. human cells. J. Cell Sci. 108, 635-644. Roussel, P., Andre, C., Masson, C., Géraud, G. and Hernandez-Verdun, D. Hozier, J., Renz, M. and Nehls, P. (1977). The chromosome fibre: evidence (1993). Localization of the RNA polymerase I transcription factor hUBF for an ordered superstructure of nucleosomes. Chromosoma 62, 301-317. during the cell cycle. J. Cell Sci. 104, 327-337. Igó-Kemenes, T. and Zachau, H. G. (1977). Domains in chromatin structure. Saitoh, N., Goldberg, I. G., Wood, E. R. and Earnshaw, W. C. (1994). ScII: Cold Spring Harb. Symp. Quant. Biol. 42, 109-118. an abundant chromosome scaffold protein is a member of a family of putative Jackson, D. A., McCready, S. J. and Cook, P. R. (1984). Replication and ATPases with an unusual predicted tertiary structure. J. Cell Biol. 127, 303- transcription depend on attachment of DNA to the nuclear cage. J. Cell Sci. 318. Suppl. 1, 59-79. Saitoh, Y. and Laemmli, U. K. (1994). Metaphase chromosome structure: Jackson, D. A. and Cook, P. R. (1985). Transcription occurs at a bands arise from a differential folding path of the highly AT-rich scaffold. nucleoskeleton. EMBO J. 4, 919-925. Cell 76, 609-622. Jackson, D. A. and Cook, P. R. (1988). Visualization of a filamentous Scheer, U. and Rose, K. M. (1984). Localization of RNA polymerase I in nucleoskeleton with a 23 nm axial repeat. EMBO J. 7, 3667-3677. interphase cells and mitotic chromosomes by light and electron microscopic Jackson, D. A., Yuan, J. and Cook, P. R. (1988). A gentle method for immunocytochemistry. Proc. Nat. Acad. Sci. USA 81, 1431-1435. Nuclear and chromosome structure 2935

Shermoen, A. W. and O’Farrell, P. H. (1991). Progression of the cell cycle Townes, P. L. and Holtfreter, J. (1955). Directed movements and selective through mitosis leads to abortion of nascent transcripts. Cell 67, 303-310. adhesion of embryonic amphibian cells. J. Exp. Zool. 128, 53-120. Sorsa, V. (1986). Distribution of chromomeres as a basis of chromosomal Traub, P. and Shoeman, R. L. (1994). Intermediate filament and related coiling. J. Cell Sci. 80, 193-205. proteins: potential activators of nucleosomes during transcription initiation Spector, D. L. (1993). Macromolecular domains within the cell nucleus. Annu. and elongation? BioEssays 16, 349-355. Rev. Cell. Biol. 9, 265-315. van Holde, K. E. (1989). Chromatin. Springer-Verlag, NY, USA. Steinberg, M. S. (1964). The problem of adhesive selectivity in cellular Vogelstein, B. Pardoll, D. M. and Coffey, D. S. (1981). Supercoiled loops and interactions. In Cellular Membranes in Development (ed. M. Locke), pp. eukaryotic DNA replication. Cell 22, 79-85. 321-366. Academic Press, New York. Wansink, D. G., Schul, W., van der Kraan, I., van Steensel, B., van Driel, R. Steinberg, M. S. and Takeichi, M. (1994). Experimental specification of cell and de Jong, L. (1993). Fluorescent labelling of nascent RNA reveals sorting, tissue spreading, and spatial patterning by quantitative differences in transcription by RNA polymerase II in domains scattered throughout the cadherin expression. Proc. Nat. Acad. Sci. USA 91, 206-209. nucleus. J. Cell Biol. 122, 283-293. Strätling, W. H., Muller, V. and Zentgraf, H. (1978). The higher order repeat Wilson, E. B. (1928). The Cell in Development and Heredity. 3rd edition, pp. structure of chromatin is built up of globular particles containing eight nucleosomes. Exp. Cell Res. 117, 301-311. 906-911. Macmillan, Inc., New York. Sumner, A. T. (1991). Scanning electron microscopy of mammalian Woodcock, C. L. (1994). Chromatin fibres observed in situ in frozen hydrated chromosomes from prophase to telophase. Chromosoma 100, 410-418. sections. Native fiber diameter is not correlated with nucleosomal repeat Swanson, C. P. (1942). Meiotic coiling in Tradescantia. Bot. Gaz. 103, 457- length. J. Cell Biol. 125, 11-19. 474. Yoda, K., Kitagawa, K., Masumoto, H., Muro, Y. and Okazaki, T. (1992). Swedlow, J. R., Sedat, J. W. and Agard, D. A. (1993). Multiple chromosomal A human centromere protein, CENP-B, has a DNA binding domain populations of topoisomerase II detected in vivo by time-lapse, three- containing four potential α helices at the NH2 terminus, which is separable dimensional wide-field microscopy. Cell 73, 97-108. from dimerizing activity. J. Cell Biol. 119, 1413-1427. Thoma, F., Koller, T. and Klug, A. (1979). Involvement of histone H1 in the Zatsepina, O. V., Polyakov, V. Y., Chentsov, Y. S. (1983). Chromonema and organization of chromatin and the salt dependent superstructures of chromomere: structural units of mitotic and interphase chromosomes. chromatin. J. Cell Biol. 83, 403-427. Chromosoma 88, 91-97.